The anatomy of job polarisation in the UK

Journal for Labour Market Research

Table 3 Shift-share decomposition of changes in occupational shares (pp) by type of occupations using alternative routine classifications, 1979–2012

	Non routine occupations			Routine occupations
	Total	Between	Within	Total	Between	Within
(A) Routine classification based on Acemoglu and Autor (2011)
Bottom	4.2	− 4.6	8.8	− 0.7	− 4.3	3.6
Middle	2.6	1.6	1.0	− 21.8	− 8.8	− 13.0
Top	16.6	16.3	0.3	− 0.8	− 0.1	− 0.7
All				− 23.4	− 13.2	− 10.1
(B) Routine classification based on RTI index from Goos et al. (2014)—top 30%
Bottom	6.1	− 6.9	12.9	− 1.6	− 1.1	− 0.4
Middle	− 7.9	− 2.9	− 4.9	− 11.5	− 1.4	− 10.1
Top	14.9	12.3	2.6	0.0	0.0	0.0
All				− 12.7	− 3.4	− 9.3
(C) Routine classification based on RTI index from Akçomak et al. (2016)—top 30%^a
Bottom	6.5	− 4.2	10.7	− 3.0	− 4.7	1.7
Middle	− 6.7	− 1.5	− 5.1	− 12.6	− 5.7	− 6.9
Top	15.9	16.2	− 0.3	− 0.1	0.0	− 0.1
All				− 15.7	− 10.5	− 5.3

Results from a shift-share analysis with 48 groups: 4 education groups, 3 age groups, gender, immigration status. Details on the routine classifications are provided in Sect. 6 and Additional file 1: Appendix S4
The discrepancies between the totals in Panel B and the other two panels is due to the fact that the RTI index from Goos et al. (2014) is only available for 21 ISCO 88 codes
^aDue to the size of the underlying occupations, the actual initial share of routine occupations here is 40% as shown in Table 3